Publishers of web-based content may determine optimal configurations for presenting content on their web pages, for example, using multivariate testing techniques, to encourage certain user actions. Measuring such user actions, however, may be problematic when the actions are dissociated from the publisher of the content. For example, a publisher may not be aware of which configuration of its web page led to a user completing an online survey after clicking on an advertisement (“ad”) on the web page associated with a third party advertiser. Publishers, therefore, may be unable to attribute certain user actions to their web page configurations, and therefore, may be unable to measure the effectiveness of their content presentations.
Certain implementations will now be described more fully below with reference to the accompanying drawings, in which various implementations and/or aspects are shown. However, various aspects may be implemented in many different forms and should not be construed as limited to the implementations set forth herein; rather, these implementations are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. Like numbers refer to like elements throughout.